JP3230767B2 - Method and apparatus for transmitting compressed image signal using trellis coded QAM - Google Patents

Abstract

Coded modulation schemes based on codes for QPSK modulation are directly incorporated into QAM based modulation systems, forming trellis coded QAM, to provide a practical coding structure that is both efficient in bandwidth and data reliability. Concatenated coding with QPSK based trellis coding and symbol error correcting coding is used. In an encoder (Fig. 2), an N-bit QAM constellation pattern (80) is divided into four subsets, each including N/4 symbol points of the constellation pattern. A two-bit QPSK codeword (92) is assigned to each of the four subsets (82, 84, 86, 88). A symbol to be transmitted is first encoded using an outer error correcting encoding algorithm, such as a Reed-Solomon code (12). Part of the symbol is then encoded (48) with an inner code that comprises a rate 1/2 trellis encoding algorithm to provide a QPSK codeword, which is mapped (50) with the remaining bits of the symbol to provide a modulation function, wherein the remaining bits (94) correlate the symbol with one of the symbol points included in the subset defined by the QPSK codeword. At a receiver (Fig. 3), the recovered modulation function is pruned (62) to provide a set of metrics (66) corresponding to the subsets and to provide a plurality of conditional determinations of the constellation point identified by the remaining bits. The metrics are used in a rate 1/2 trellis decoder (68) to recover a first bit that is encoded using a rate 1/2 encoding algorithm to recreate the QPSK codeword. One of a plurality of the conditional determinations is selected in response to the recreated codeword. The selected conditional determination is combined with the recovered first bit to provide a decoded output that is further decoded using a symbol error correcting algorithm such as a Reed-Solomon code (36). <IMAGE>

Description

DETAILED DESCRIPTION OF THE INVENTION [0001] The present invention was filed on July 26, 1991.
Of U.S. Patent Application No. 07 / 736,738
It is a continuation application. [0002] BACKGROUND OF THE INVENTION The present invention relates to a trellis-coded orthogonal
Regarding width modulation (QAM), more specifically, QAM transmission
It relates to a practical method of encoding the transmission. The present invention, in particular,
Pressure in high definition television (HDTV) systems
The present invention can be applied to transmission of a reduced image signal. [0003] 2. Description of the Related Art For example, high definition television (HDT)
V) For each digitized image signal used for signal broadcasting
Digital data is used for VHF and UHF ground communication
It can be transmitted to the end user via the analog transmission path.
In analog transmission lines, the input waveform is distorted or deformed.
Transmitted. Waveform distortion is usually statistical
However, dark noise, impact noise, fading, etc. may occur
As such, they may be additive and / or synergistic. Biography
The deformation caused by the transmission path is frequency conversion, nonlinear
Or harmonic distortion and time dispersion. [0004] Digital data is transmitted through an analog transmission path
This can be done, for example, using pulse amplitude modulation (PAM).
Modulate data. Generally, quadrature amplitude modulation (QAM)
Data that can be transmitted within the bandwidth of the available channel.
Increase the amount of data. QAM is a type of PAM,
The number of bits information contains, for example, 16 or 32 points
One pattern called "constellation"
Are transmitted together. In the case of pulse amplitude modulation, each signal is transmitted
Pulse with an amplitude level determined by the
You. 16pointIn QAM, typically each quadrature channel is-
Use symbol amplitudes of 3, -1, 1 and 3. Ma
In 32-bit QAM, generally, -5, -3, -1,
Use 1, 3, and 5 symbol amplitudes. Digital
The bandwidth efficiency of a telecommunications system is the transmission rate per second per unit bandwidth.
It is defined as the number of transmission bits. That is, the data for the bandwidth
It is the ratio of tarates. High data rate and small
In applications where occupied bandwidth is required, bandwidth efficient
A tuning system is used. QAM is a band-efficient modulation
It is. [0006] On the other hand, four types commonly found in satellite communication systems
Modulation methods such as phase shift keying (QPSK)
Well established and understood. QPSK is better than QAM
A simple constellation pattern is obtained. In particular,
Constellation pattern used in QPSK system
Are generally equal in phase with a phase difference of 90 degrees from each other.
It contains only four symbols of width. Therefore, four
The symbols are evenly spaced around the circle. In QPSK modulation, band limitation is a major problem.
Suitable for systems with no but limited output. one
On the other hand, in QAM modulation, the output requirement is not a major issue
Is advantageous in systems where is limited. Therefore, QP
SK is used as an excellent method in satellite communication systems.
Has been used. On the other hand, QAM is for ground and cable systems.
Suitable for systems. As a result of the spread of QPSK, trellis marks
An integrated circuit that realizes encrypted QPSK modulation has already been made.
And can be easily obtained. [0008] Trellis coded modulation (TCM) is a
Coding and modulation for digital transmission over limited channels
It has evolved as a combined technology. This is the QAM
Compared to conventional uncoded multi-level modulation,
Deliver important coding benefits without compromising bandwidth efficiency
Can be achieved. The TCM scheme uses redundant non-binary modulation,
Governs the choice of modulated signal and generates coded signal sequence
It is used in combination with a finite state encoder. On the receiving side
Is used to convert a noisy signal to a soft-decision maximum likelihood sequence decoder.
To decrypt. These schemes provide additive noise
The robustness of digital transmission has been
Tone can be increased by 3-6 dB. like this
A significant advantage is the bandwidth extension required by other known error correction schemes.
Achieved without significant or reduced effective information rates.
The term "trellis" is used for binary convolution
State transition (trellis) diagram similar to the trellis diagram of
This is because this method can be explained. the difference
Is based on the principle of convolutional coding,
This is to extend to non-binary modulation by a set. For realizing trellis coded QPSK modulation,
Availability of components for QP
For applications such as satellite communications where SK technology is dominant
This is a great advantage for designing low-cost communication systems.
You. However, these components make QAM
To realize other coded transmission systems such as
Did not stand. [0010] Both output and bandwidth are limited and low cost
Required components (especially low-cost data decoders)
In the required applications, traditional QAM systems are necessary
Encoder and decoder circuits are complex and relatively expensive
It seemed to be useful because of its value. In fact, QA
The M trellis encoder and decoder is a costly custom collection.
It is generally realized in an integrated circuit chip. [0011] The output and bandwidth are limited and
Digital data communication needs low-cost solution 1
One application area is for compressed high-definition television signals.
There is digital communication. A system for transmitting compressed HDTV signals
Stem is about 15-20 megabits per second (Mbps)
Data rate, occupied band of about 5-6 MHz (conventional
According to the National Television Standards Committee (NTSC)
TV channel bandwidth), and very high data
Reliability (that is, very low bit error rate)
You. This data rate requirement is for high quality compressed television images
Arising from the need to provide. The bandwidth constraint is H
DTV signals use existing 6 MHz television channels
And must coexist with the current NTSC broadcast signal
Not by the U.S. Federal Communications Commission.
You. Full HDTV performance in a single 6MHz band
High-efficiency proprietary based on DCT variant coding
The compression algorithm isIEEE minutes for broadcastThird
Vol. 6, No. 4, pp. 245-254 (December 1990)
W. Pike's Digicipher-Fully Digital Communications
Road Compatible HDTV Broadcasting System "(W. Paik," Digiciph
er-All Digital, Channel Compatible, HDTV Broadca
st System ",IEEE Transactions on Broadcasting, Vo
l. 36, No. 4, December 1990, pp. 245-254)
Has been proposed and is incorporated herein by reference. With the combination of the data rate and the occupied bandwidth,
Requires a modulation system with high bandwidth efficiency. In fact,
Data rate to bandwidth ratio should be around 3 or 4
No. That is, a QP having no bandwidth and a band efficiency of 2
A modulation scheme such as SK is not suitable. Higher bandwidth efficiency
A modulation scheme such as QAM is required. But first
As mentioned, QAM systems are used in high-consumption applications.
The difficulty was that it was too expensive to be realized in the field. Very high data reliability in the HDTV field
Is required because of the highly compressed source material (that is,
(Compressed image) is vulnerable to transmission line errors. De
To concisely describe the original value of the data,
Length has been excluded. For example, one error every 24 hours
In the case of a system transmitting at 15 Mbps as less than
The bit error rate (BER) is 10¹²Number of transmission bits
Error must be less than one time. The data reliability requirements are divided and cooperated (divi
de and concur) to solve the problem
The use of techniques often results in
No. In such a coding framework, two codes are used.
The "inside" modulation code cleans the transmission path and the "outside"
Transmit a reasonable symbol error rate to the decoder. The inner code is
Usually, "soft decisions" (that is, elaborately quantized transmission
Coded modulation that can be effectively decoded using
You. It is well known that the inner code is convolution or
A form of "Viterbi algorithm" using Leris code
Is a trellis decoder. The outer code is almost
In each case, the t-error correcting "Reed-Solomon" code
is there. Data rate range required for HDTV data communication
Such a Reed-Solomon encoding system operating in a
The system is widely available and can be integrated by several manufacturers.
Is realized on the road. Outer decoder passes through inner decoder
Removes the majority of symbol errors
The force error rate becomes very small. For a more detailed description of the concatenated coding method,
G. FIG. C. Clark Jr., J.M. B. Kane co-authored "De
Error Correction Coding for Digital Communications "(1981) (G.C.
Clark, Jr. and J.B. Cain, "Error-Correction Coding
for Digital Communications ", Plenum Press, New Yor
k, 1981); Lin, D. J. Costello Ju
Nia, "Error Control Coding: Basics and Applications" (1983)
Year) (S. Lin and D.J. Costello, Jr., "Error Contro
l Coding: Fundamentals and Applications ", Prentice
-Hall, Englewood Cliffs, New Jersey, 1983)
be able to. For trellis coding,information
IEEE conference proceedings on theoryIT-28 Vol. 1, No. 55
On page 67 (January 1982). By Ungerbeck
"Transmission Channel Coding Using Multilevel / Phase Signals" (G.U.
ngerboeck, "Channel Coding with Multilevel / Phase S
ignals ",IEEE Transactions on Information Theory,
Vol. IT-28, No. 1, pp. 55-67, January 1982),I
EEE Communication MagazineVol. 25, No. 2, pp. 5-21 (1987
February) G. "Using a redundant signal set"
Trellis Coded Modulation-Part I: Introduction, Part II:
Latest Technology "(G. Ungerboeck," Trellis-Coded Modulati
on with Redundunt Signal Sets-Part I: Introduct
ion,-Part II: State of the Art ",IEEE Communica
tions Magazine, Vol. 25, No. 2, pp. 5-21, February
 1987), andIEEE Conference Proceedings on Information TheoryI
T-33 Vol. 2, No. 177-195 (March 1987)
A. R. Calderbank, N.M. J. A. Sloan
"New Trellis Codes Based on Lattice and Coset Systems" (A.R.
 Caulderbank and N.J.A.Sloane, "New Trellis Codes
 Basedon Lattices and Cosets ",IEEE Transactions o
n Information Theory, Vol.IT-33, No. 2, pp. 177-19
5, March 1987). Viterbials
For the gorhythm,IEEE BulletinVolume 61 Issue 3
G. (March 1973). D. By Forney Jr.
"Viterbi Algorithm" (G.D. Forney, Jr., "TheVite
rbi Algorithm ",Proceedings of the IEEE , Vol. 61,
 No. 3, March1973). Lead-Solo
For the Mont coding scheme, see Clark Jr.
(Clark, Jr. et al) and Lin et al.
Is discussed in the article. [0016] SUMMARY OF THE INVENTION
Error rate performance at the output of the inner modulation code
It largely depends on the noise-to-noise ratio (SNR). Some codes are SN
The performance is good when R is low, the error rate is low,
The signal performs better when the SNR is high. That is, concatenation
Modulation codes for both coding and non-concatenated coding systems
Different solutions for different SNR ranges
It means that you need it. In the HDTV broadcasting system, the coverage
(Effective range) / Characteristics where station spacing and video quality conflict with each other
You need to find the best balance
is there. Low-order QAM (eg, 16-QAM)
Due to the performance characteristics of low wave-to-noise ratio, high-order QAM
Provide higher coverage (eg, 64-QAM),
Station intervals can be made tighter. On the other hand, higher QAM
Has higher image quality than low-order QAM
provide. Which order of QAM to choose depends on the terrain
And available / authorized transmitter outputs and channels
Very often depends on the conditions of these
Parameters can very often be determined by the sender
QAM that can automatically select the QAM transmission mode
It can be equipped with a communication system. Such a cis
The system will of course automatically determine the order of the QAM used at the transmitter.
Equipped with a receiver that can detect with high reliability
You must be able to receive the signal correctly. This
A system that provides these features is described in March 24, 1992.
U.S. Patent issued as commonly assigned co-pending application
Application No. 07 / 852,330 "Mode-selective orthogonal vibration
Width Modulation Communication System ", by reference
Incorporate here. For communication of HDTV signals including compressed images
Data modulation systems with high bandwidth efficiency and low output requirements
It would be advantageous to equip the system. Such a system
High data with minimum occupied bandwidth and very high data reliability
Must have tallate. Mass production at low price
The complexity of the receiver used in these systems
Must be minimized. The system is individualized
Components as easily as possible
Must be able to be realized using The present invention provides a modulation system having the above advantages.
I will provide a. In particular, the method and apparatus of the present invention provide
Trellis coded QPSK system without sacrificing reliability
The range of the stem is trellis mark especially useful for HDTV communication.
Extends to encrypted QAM systems. [0020] According to the present invention, high definition is provided.
A method for transmitting a television signal is provided. HDT
The image portion of the V signal is divided into PCM video data blocks.
Crack. Process the block using motion prediction and compensation,
Generate DPCM data. Based on predetermined criteria
PCM image data and DPCM for each block
One of the data is selected for transmission. For example, send
Choose an alternative that produces the least number of bits for
Can also be. The selected data is transformed using the discrete cosine transform
To generate a block of transform coefficients. Then this
Is quantized to increase its coding efficiency. Quantized transformation
The permutation coefficients are then variable-length coded. Resulting
Coefficients are represented by Reed-Solomon outer code and trellis inner code
For encoding and transmission using an encoding structure concatenated with
Is generated. Code using quadrature amplitude modulation
The coded signal. In the preferred embodiment, the Reed-Solomon outer code
Symbols generated by the first interleaving
Interleave according to the
Bursts that can be caused by the Leris inner code
Disperse errors. The sign produced by the trellis inner code
The encoded signal according to the second interleave format
Interleaved, and then to the encoded signal transmission path
Distribute burst errors that may occur along the
You. It contains compressed image information and uses quadrature amplitude modulation.
In the case of a high-definition television signal transmitted by
Receiving a carrier that includes a signal. Demodulates the received carrier
To restore the interleaved modulation function containing the compressed image information
I do. The reconstructed modulation function is then deinterleaved,
Decode with inner trellis decoding algorithm of concatenated decoder
Interleaved leads representing compressed image information
-Restore Solomon symbols. Restored lead-source
De-interleave the Romon symbols outside the concatenated decoder
Input to the Side Reed-Solomon decoding algorithm. Lee
The De-Solomon decoding algorithm is deinterleaved.
Variable-length coding coefficient from the obtained Reed-Solomon symbol
Restore. Decode the restored variable-length coded coefficients, and
The transform coefficient representing the image information is restored. Next, the conversion coefficient
Is inversely converted to the PCM and DPCM formats.
At least one image data is restored. Restored
DPCM data processed using motion compensation and restored
The PCM image data represented by DPCM is restored. Return
Source PCM image data to an HDTV
Format for output to the transceiver. The present invention further provides a specific HDTV signal.
It also provides a QAM transmission structure. QAM controller consisting of N points
Divide the stellation pattern into four subsets
You. Each subset has N constellation patterns
/ 4 symbol points. Each of the four subsets
Are assigned different 2-bit codewords. Transmission
Is converted to a binary convolutional coding algorithm with a coding rate of 1/2.
By processing the first bit of the symbol
Encoded in the constellation pattern.
2-bit code assigned to the subset in which the symbol exists
Generate a word. This 2-bit codeword is used as the remainder of the symbol.
And a modulation function to generate a modulation function. remaining
Bits are included in the subset defined by the codeword.
One of the N / 4 symbol points
It is something to make. Change the carrier for transmission over the communication path.
Modulate with a tonal function. In the embodiment shown, the 2-bit codeword is
Construct the least significant bit of the modulation function
Defines a column in the matrix of pattern coordinates. Remaining
The bits make up the most significant bit of the modulation function,
Determine the size of the translation pattern. Place of connection method
If the information bits are first, for example, Reed-Solomon codes
Using a t-symbol error correction code such as
Encoding. These encoded symbols are then
To the trellis encoder, where the desired modulation on the carrier is
Do. After transmitting the modulation function, it is restored on the receiving side.
I will. Pruning the reconstructed modulation function to a subset
Generate a corresponding set of metrics and leave the remaining bits
Display the various conditional decisions of the signal points identified by
Generates many bytes to pass. This metric is
Algorithm for Decoding Binary Convolutional Code with Half Rate
(E.g., the Viterbi algorithm)
Restore The reconstructed first bit is calculated by coding rate 1 /
2 used in the binary convolutional encoding algorithm
To play. In response to the reconstructed codeword, a conditional decision
Select one of the constant bytes. Next, the selected byte is
Decoded output in combination with the first recovered bit
Generate The present invention further provides a DAM for QAM transmission.
An apparatus for encoding digital data is provided. The encoder is
The symbol to be transmitted is identified by the first bit and at least one
And parsing means for decomposing the bits into bits. Also,
Binary convolutional coding algorithm with coding rate 1/2
Encoded in rhythm, NpointQAM constellation par
2-bit code defining one of the four subsets of turns
Equipped with means to create words. Each subset is
N / 4 symbol points of the stellation pattern
PiecesWithoutIncluding one. Map the codeword with the remaining bits and modulate
Generate a function. Remaining bits defined by codeword
1 of N / 4 symbol points included in the subset
One to the symbol. Also, when transmitting over a communication path,
For this purpose, a means for modulating a carrier with a modulation function is provided. Further
The information bits using an error correction algorithm
To generate symbols that are decomposed by
Outer encoders can be provided. In the embodiment shown, the codeword is the modulation function
Forming the least significant bit, the constellation pattern
Define the columns of the coordinate matrix of the button. Remaining bits
Constitutes the most significant bit of the modulation function and constellation
Determine the size of the action pattern. Encoding means trellis
A coding algorithm can be used. A decoding device is also provided according to the invention. Receiving
Machine demodulates the received carrier andpointQAM modulation function
To restore. In this modulation function, a 2-bit codeword
Assigns one of many QAM constellation subsets
Identify and rest (nThe bit portion of -2) is the one sub-section.
Represents one signal point in the set. In addition, the restored changes
Pruning the tonal function and matching the subset
Generate a corresponding set of metrics, and (n-2) bit
Multiple conditional decisions of signal points identified by
Multiple (n-2) Create subgroup of bits
Equipped with pruning means. This metric is encoded
Used to decode binary convolutional codes with a rate of 1/2.
To restore the first bit. Next,
The first bit is a binary convolutional coding algorithm with a coding rate of 1/2.
Encode using a algorithm and reproduce the codeword. Further
In response to the reproduced codeword, a number of (n-2)
Equipped with a means to select one from a subgroup of
You. First bit restored selected subgroup
To generate a decrypted output. In the embodiment shown, the codeword is the modulation function
Consists of the least significant bit,
Rix columns, while selected subgroups
Constitutes the most significant bit and defines the row of the matrix
You. The pruning means uses a matrix of constellation coordinates.
Of each columnpointQuantizes the modulation function
The conditional decision value consists of an optimal selection for each column,
The set of tricks represents the quality of each choice. This metric
And decode the convolutional code using the soft decision algorithm
Use with decoders A concatenated decoder is also provided. Example of connection method
Now, use the error correction algorithm to decode the output.
Equipped with side decoder. In the embodiment shown, the concatenated decoder
The inner decoding algorithm used for
Consists of The outer symbol error correction algorithm is
It can be composed of a code-Solomon code. At the receiver
The received carrier signal is a high definition television carrier
It can be composed of signals. [0031] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows a connection for transmitting QAM data.
1 shows an encoding system of the scheme. Digital information to be transmitted
Is connected to the input terminal 10 through a system such as a Reed-Solomon encoder.
It is input to the symbol error correction encoder 12. The encoder 12
The information is composed of a number of consecutive n-bit symbols 16.
Is converted into a code word 14. Outer convolutional code for encoder 12
Can be used, but errors can be concentrated in the transmission system.
Easy to use, only hard quantized data can be used
The fact that fast codes are desirable.
Then consists of an n-bit segment of the binary stream
Reed-Solomon code using symbols as outer code
That's the right choice. The Reed-Solomon code is
Since it depends only on the number of symbol errors in the block, n
Susceptible to burst errors within one symbol of bits
Absent. However, the performance of the concatenated system is
Degraded severely by long bursts. Accordingly
To the output of the Reed-Solomon encoder 12
The encoder 18 is installed and during the encoding operation (for individual bits
No) to insert the symbol. Purpose of pinching
Is to decompose the burst of symbol errors. The interposed symbol is a QAM trellis code.
The signal is input to the signal generator 20. In the present invention, the encoder 20
As described in detail below, QPSK codes are trellis codes
Embedded QAM modulation system. The output of the encoder 20 is a QAM constellation
The coordinates of the real plane (I) and the imaginary plane (Q) of the
Consists of symbols that represent One such constellation
The action point 22 is symbolically shown in FIG. This symbol
Transmitted by a conventional transmitter 24 via a transmission line 26.
It is. The transmission path introduces various distortions and delays, and the signal
Distorted before being received by receiver 28. As a result,
The coordinate value realized for the symbol is the transmitted coordinate value.
Is not strictly correlated with the received point 30
It differs from the actually transmitted point 22 of the terrain pattern.
Position. Determine the exact location of the receiving point and
Therefore, in order to obtain the data as actually transmitted,
QAM trellis using decision convolution decoding algorithm
The received data of the following equation 4 is input to the decoder 32,
Restore information. The decoder according to the invention is described in more detail below.
I do. (Equation 4)The output decoded by the decoder 32 is
Deinterleave which has the opposite effect to interleaver 18
Is input to the server 34. Deinterleaved data
Is input to the Reed-Solomon decoder 36 and the original information
Restore bits. In the present invention, a QPSK code is a trellis code.
Low SNR area by incorporating
Low bit error rate and high data rate when
To provide a band efficient system. Achieve this goal
QPSK code that defines one symbol for both
The code word of the signal and the “uncoded” bit are uniquely
Assign to stellations. Also, a soft decision decoder and
Determine the constellation point represented by the "uncoded" bit
Decoding the received signal by combining the
You. FIG. 2 shows an encoder according to the invention. Day
Tabits (eg, the output of the interleaver in FIG. 1)
Input from the input terminal 40 to the conventional purging circuit 42.
You. SendnGenerated by decomposing a 1-bit symbol
The first bit is output to line 46 and
Sent to the container 48. Remainingn-2 "uncoded" bits
Output to line 442 ⁿ To the QAM mapping circuit 50
Sent. The convolutional encoder 48 converts the generator in octal
171 and 133, coding rate 1/2, number of states 64
Is used. 2-bit output of encoder 48
When(n-2) The uncoded bits are2 ⁿ -QAM photo
Sent to the image circuit,nQAM Constellation of Bit Symbols
A specific constellation point on the constellation
Used as a label for From convolutional encoder 48
The "encoded" 2-bit output of is actually QPSK
A codeword that can be used to select a constellation subset.
used. The uncoded bits are the QAM constellation
From a specific signal in the constellation subset
Used to select points. In QAM transmission (encoding), a QPSK code
Codeword and the remaining uncoded bits in the QAM
Must be assigned to the For this purpose
The following modulation function MOD (m) （R^TwoBy QAM
The labeling of the constellation points must be described.
No. MOD: {0,1}^N→ R^Two The mapping described below has the following desirable features:
ing. (1) The influence of the ambiguity of the quadrature of QAM is QP
Appears in the SK codeword, but the uncoded bits are a shadow of ambiguity.
Unaffected (ie, the quadrature ambiguity is
Treated in the same way as stems). (2) Highest day
Git controls the size of the constellation.
Nested structure of 16/32 / 64-QAM). 16-QAM (m_Five= M_Four= 0), then
Consider the labeling described by the matrix of
(M_Five= M_Four=m _Three =m _Four QPSK if = 0). (Equation 1) 32-QAM (m_Five= 0), the following matrix is
to add. (Equation 2)For 64-QAM, add the following matrix: (Equation 3) The output of the QPSK encoder is the highest of the modulator inputs.
Lower bit (LSB) m₁m₀Make up the matrix
Select a column. Most significant bit (MSB) is constellation
Determine the size of the application. No uncoded bits (m
_Five= M_Four= M3 = m4 = 0), QPSK is generated
You. When there are two uncoded bits (m_Threem_Two), 16-Q
An AM is generated. When there are three uncoded bits (m_Four
m_Threem_Two), 32-QAM is generated. Uncoded bits
Is four (m_Fivem_Fourm_Threem_Two), 64-QAM generated
It is. Otherwise, the QAM constellation is 9
Due to the effect of rotating by 0 degrees, the columns of the matrix
To rotate. 00 → 01 → 11 → 10 → 00; The line does not change. That is, labeling uncoded bits
Is equivalent to 0 °, 90 °, 180 °, and 270 ° rotation
Therefore it is not affected. Right angles at the receiver (decoder)
The handling of phase ambiguity is left solely to the QPSK encoder.
You. The method used to resolve ambiguity in QPSK receivers is
QAM system using this labeling in any way
Can be directly incorporated into For example, the QPSK code itself
When the body does not change cyclically, the differential encoding of QPSK is used.
Can be used. 16-QAM and 32-Q according to the invention
Label the AM constellation pattern
Figure 4 shows a diagram. Constellation generally indicated by numeral 80
The pattern of the 16-QAM and 32-
It corresponds to the matrix of QAM. Specifically, 16-QA
To explain the example of M, 16 dashes are included in a dashed square 90.
There is a constellation point. These constellations
The points are identified by tokens 82, 84, 86, 88 in FIG.
Are divided into four subsets as shown in FIG. Each subset
The set includes four constellation points. Thus, white
In the case of the subset 82 shown by a circle, four
Points 82a, 82b, 82c, and 82d. Sa
The subset itself is composed of two encoded video elements, indicated by numeral 92 in FIG.
(QPSK bits) defined by m0 and m1
You. For 16-QAM, a specific point in each subset
Are the "uncoded" bits m2, m shown by numeral 94 in FIG.
3. Thus, 82c is a subset
00, defined as point 011 in this subset.
84a, 86a and 88a etc.
Option points are similarly identified. In the case of 32-bit QAM, a dashed square 90
16 points outside of are added. These points as well
Labeled, in this case the three shown at 94 in FIG.
, M3, m4 are used. like this
Labeling can be extended to higher levels of QAM
It will be understood. Features of the labeling structure used in the present invention
Is the Hamming of each QPSK symbol as shown in FIG.
Weight is equal to Euclidean weight divided by factor x
That's what it means. Where x is the constellation point
Between (minimum distance)^TwoCorresponding to In the embodiment, shown in FIG.
So that on each quadrature channel
Consume to QAM levels of 1, -1, 3, -3, 5, and -5
Because there are terrain points,
The minimum distance is 2, and the Hamming weight is Euclidean weight
Divided by 4. FIG. 3 shows a QAM trellis demodulation according to the present invention.
Shows the implementation of the vessel. The received symbol data is
60 is input to the pruning circuit 62. The pruning circuit 62 is
Process the derived modulation function and use the QPSK codeword
Generate a set of metrics corresponding to a defined subset
As well as identified by the transmitted uncoded bits.
Multiple (representing multiple conditional decisions ofn−
2) Create a subgroup of bits. In particular, the four
The trick outputs to line 66, the coding rate is 1/2, the number of states
64 Viterbi decoder 68. Four sets of (n-2)
The conditional determination of the bit is output on line 64. The pruning circuit 62 includes a pre-calculated metric
(Equation 4)
A program that stores a reference table containing conditional judgment
Storage device such as a programmable read only memory (PROM)
Configuration. The value that can be expressed by (Equation 4) is
Addressing the PROM, the corresponding stored method
Used to output licks and decisions. to this
Thus, an ultra-high-speed pruning operation can be performed. Viterbi decoder
Using the accumulation of metric history received from the pruning circuit
To decode the QPSK codeword. The Viterbi decoder 68 shown in FIG.
Conventional coding rates 1/1 available for PSK coding configurations
2 decoders. Therefore, the present invention
Use a custom-ordered Viterbi decoder to implement the decoder
It is not necessary to decode the trellis code using the method. A system using the QAM modulator described above.
Signal detection when a soft decision QPSK decoder is incorporated in the system
Think of the process. First, a QPSK or QAM signal
In the hard decision detection process of_k= X_k+ W_kQuantize
I do. Where the signal x_kIs QPSK or QAM cons
Belongs to the teration (that is, the range of MOD (m)), w
_KIs noise. The quantization function is expressed by the following equation (7).
Therefore, the signal shown by the following (Equation 5) and the signal shown by the following (Equation 6)
Estimate both data. For maximum likelihood search (ML),
Possible message m {0,1}^NLog likelihood function in
Number-log (p (y_k│MOD (m)) to a minimum. (Equation 5) (Equation 6) (Equation 7) Where p (y_k | x_k) Is x_kWhen you send y_kTo
The conditional probability of receiving. A place for random messages
In this case, the probability of an error is minimized by the ML detection. Likely general
A typical quantization method is nearest neighbor (Euclidean) detection.
This is (Equation 8) Meet. However, ‖ ・ ‖^TwoIs the square of the Euclidean distance
(That is, the sum of squares). Field of additive Gaussian noise
In this case, the closest contact detection is ML. In coded QPSK and QAM systems
Recovers the soft decision information to effectively decode the codeword.
No. must be provided. This soft decision information is often
Often described as a symbol metric. This met
Rick is shown in (Equation 14) below when yk is received.
Indicates the quality of the determination that a particular symbol has been transmitted. [Equation 14] For nearest neighbor decoding, the metric selected is Metric (y_kM) = ‖y_k−MOD (m) ‖^Two Becomes In fact, the metric itself for implementation purposes
Are quantized. For example, each possible message m₁, M₀
{0,1}^TwoFor QPSK,
Cooky_k −MOD (m₁, M₀) ‖^TwoIs additive Gaussian
6 is an ML metric of sexual noise. Based on QPSK codes capable of soft decision decoding
In Leris coded QAM modulation, four symbol metrics
And provide four conditional hard decisions to the decoder
Must. M shown in the following (Equation 9)₁, M₀∈ ｛0,
1｝^TwoClosest point for each choice ofIn detection, conditionally hard
The decision is to get the minimum value _n-1 ,. . . , M _Two Supports selection
Do. (Equation 9) Symbol metric and conditional hard decision
Roses is known as pruning. Trellis coding
In QAM, uncoded bits are trellis "parallel"
Appear as a symbol and calculate symbol metrics and conditions.
By conditional hard decision, all except the single optimal branch,
Pruning from a set of parallel edges. The pruning is performed by the QAM modulation matrix described above.
Can be easily explained by The pruning operation is matri
Received symbol y in each column of_kJust including the quantization of
is there. Then the conditional hard decision becomes the optimal choice for each column,
The metric corresponds to the quality of the decision. When the pruning operation is completed, the soft decision information becomes Q
It is input to the PSK code decoder. (During this time,
The hard decision with condition is stored, and the process waits for the QPSK decision. ) QP
The SK decoder uses the soft decision information to generate QPSK information (ie,
m₁, M₀s) is decrypted. Then the remaining information (ie
m _n-1  ,. . . , M_Twos) is the decoded QPSK information
And using the previously stored conditional hard decisions,
Determined in a known manner. If the QPSK decoder is ML (QPSK
Modulation), the above pruning / QPSK decoding method is also M
L. For example, if the QPSK code is
Tabi) If it is a binary convolutional code by decoding, Q
AM trellis decoding algorithm is also the closest
(That is, search for the code word closest to the received sequence.
). In the embodiment shown in FIG.
The output metrics are decoded by decoder 68.
And the single bit output on line 46 of the encoder of FIG.
Restore the corresponding single bit. This bit has a code rate of 1
/ 2, convolutional encoder 70 with 64 states (encoder 4 in FIG. 2)
8) and a 2-bit QPSK codeword
Reproduce. Reconstructed codewords have subgroups of decoder
Delay bar by an amount of time equal to the delay introduced by C.68.
Output from the pruning circuit after being delayed by the
Four (n-2) Select one of the bit subgroups
Used to do. chosen(n-2) Bit size
The group is then converted by the parallel / serial conversion circuit 76 to the decoder 68.
And the trellis decoded output
Generate As described in connection with FIG.
The power shows a reasonable symbol error rate, which is
Must be further improved by: in this way,
The decoded output is deinterleaver 38 and Reed-Solomo
Further processing by the outer decoder 36 (FIG. 1)
Restore information bits. T—arbitrary input symbol for error correction
Read the output bit error rate estimate for the error rate
Solomon codes can be easily calculated. q = 2^l
(Extended) Reed-Solomon code in the finite field of
Parameter (n_RSk, t). Where block length n
_RS≦ q + 1, dimension k = n_RS-2t, error correction capability is wrong
t. Input symbol error rate P_inMemoryless thin with
For a bol error channel, the output symbol error rate is
(Equation 10) (Equation 10)Can be limited by Next, the following equation (Equation 11) [Equation 11] To approximate the output bit error rate. Furthermore, Lee
The l-bit symbol of the De-Solomon code is smaller
An n-bit symbol (eg, trellis coded QAM transform)
), The input error rate is
The following equation (Equation 12) (Equation 12) Becomes Where P_modIs the error rate of an n-bit symbol
It is. When using coded modulation, a "memoryless" channel
In order to guarantee, the use of pinching is essential. FIG. 7 shows a conventional code having a coding rate of 2/3.
And trellis coding according to the present invention.
When using QPSK configuration with QAM coding rate 1/2
2 shows the performance of the two coupling systems. The graph of FIG.
Is the ratio of the received signal to the carrier-to-signal ratio (CNR).
This represents the De-Solomon block error rate. One bu
Lock causes one or more m-bit symbols to go into error state
Then, a block error (that is, a codeword error) occurs. Song
Line 100 is coded 1/2, state number 64 according to the present invention.
Concatenated Reed-Solomon Trellis coded 16-QAM system
It shows the performance of the system. Curve 104 is trellis coded 32
-Represents the performance of a similar system using QAM. curve
102 is a conventional trellis with 2/3 coding and 16 states
9 represents the performance of an encoded 16-QAM decoder. Curve 106
Is a conventional trellis coding with 2/3 coding and 16 states
Represents the performance of a 32-QAM decoder. The curve in FIG. 7 shows a trellis coding simulation.
M-bit read-solomon
Incorrect P_RSsymAfter estimating the probability of
(Equation 13) (Equation 13) Calculate the probability of Reed-Solomon block error according to
It was determined by Where L is Lee
De-Solomon block length (m bits per block)
Symbol number), and t is correctable per block.
Reed-Solomon symbol error count. 16-QAM
The system uses 8-bit symbols at 116 per block.
And the 32-QAM system uses 8-bit symbols
Are used per block. Which Lead-
Solomon codes are also 8-bit Reed-Solomon symbols
Can be corrected up to five per block. The curve in FIG. 7 shows that the system
If you want or need to work below
Shows the train according to the invention represented by curves 100 and 104.
It shows that the squirrel coding method is clearly the right choice.
However, even when the CNR is high, the trellis decoder is
Using QPSK Viterbi decoding circuit chip
Can be manufactured with high trellis decoder
The inventive trellis coding method is an excellent choice. FIG. 8 shows a digital HDTV communication system.
2 shows the basic components. HDTV encoder 110 is
Image information, audio information,
Receive data and text. Encoded information
Transforms a high frequency carrier using QAM according to the present invention.
Transmitted by the transmitter VHF / UHF transmitter 114
You. In consumer's home, HDTV receiver is QAM modulated
Receive the data stream. By tuner 116
Viewers can select specific programs to watch.
You. The selected program is decoded by the HDTV decoder 118.
The HDTV monitor 120 outputs the output image signal to the audio
The audio signal is sent to the speaker 124 via the width unit 122.
Output each. Data and text are also monitored
0 to the viewer. VHF and UHF ground transmission
Adaptive to overcome common multipath distortion in transmission
Equalization circuit can be equipped. The details are described below.
A forward error correction decoder is capable of transmitting virtually all received signals.
Correct random or burst errors. FIG. 9 shows an image portion of an HDTV signal before transmission.
Of a digital image encoder that can be used to encode
It is a block diagram. Analog red / green from image sources
The blue (R, G, B) input is a pre-rotation generally denoted by 130
Processed on the road. R, G, B inputs are digitized
Before being filtered and clamped by a low pass filter.
Low-pass filters provide aliasing and other spurious signals
Is designed to remove appropriately. The clamping process is
Recover proper DC levels during horizontal blanking. After analog-digital conversion, R, G, B
The signal is converted to a YUV color space. Chrominance information
Resolution has only a small effect on perceived image quality
Thus, the resolution of the luminance can be relatively reduced. U
And the chrominance component of V is 1/4 in the horizontal direction,
It is suppressed by half in the vertical direction (decimation). The decimation in the horizontal direction is, for example,
Apply digital FIR filter before sampling
Can be achieved by Horizontal interpolation
Performs zero padding in the decoder, quadrupling the gain
This is done by applying the same filtering.
Vertical decimation by 、 equals one field.
Achieved by discarding every second. The decoder
By repeating the luminance field twice,
Restore the interlaced signal. Two different files
Vertical decimation across the field
Degrades the expression somewhat, but this deterioration is easily perceived
Not a big problem. The luminance signal (Y) is obtained by the chrominance pre-processing
Bypass the road. Therefore, full resolution is maintained
You. Next, one image block at a time is
The chrominance component and the luminance component are multiplexed one by one. Next
Then, all components are subjected to the same compression processing. In the decoder
Separates the components again and interpolates the chrominance signal back
To full resolution. The image signal uses a discrete cosine transform (DCT).
Compression in two different paths. First "PCM" path
Then, the image is subjected to DCT transform at 133, and the generated coefficient is
The quantization is performed by the quantization circuit 135. Motion prediction and compensation methods
Based on the prediction of the appearance of the image frame
In the second "DPCM" path, which provides another signal, the prediction and
The difference between the actual images is DCT transformed at 134. As a result
The DCT transform coefficient obtained by the quantization is quantized by the quantization circuit 136.
And outputs the result to the selector 137, where
Based on predetermined criteria, such as whether the minimum number of bits
Quantum from either PCM or DPCM path
Select the normalized coefficient. Select each block of image data
The selected coefficient is input to the variable length encoder 138. this
Can be configured with a conventional Huffman encoder, for example.
Wear. Outputs a variable-length code word to the FIFO register 140.
And then output further to the transmitter. The DCT assigns a block of pixels to a new block.
To the conversion coefficient of the block. In the preferred embodiment, an 8 × 8
Use block size. The reason is that this size
Beyond that, the efficiency of transform coding does not improve much,
This is because the complexity becomes considerably high. Then do the conversion
Apply to each block and convert the whole image. Decoder
Now apply the inverse transform to restore the original image. DC is used to compress the energy to a small number of coefficients.
T may not be effective. For example, if the input signal is
In the case of white noise, the image energy remains in the pixel area even after conversion.
Is randomly distributed as much as when This
Under these conditions, image compression becomes much more difficult and
If you don't introduce some form of false image or something,
Can not be. Fortunately, under these conditions
Is much more noticeable than in quieter conditions
Tends to run out. In addition, such conditions are
This is not common in images, but is usually
There is a high degree of horizontal and vertical correlation. Image compression technique used in the system of the present invention
Reduces the number of bits required to represent the DCT coefficients
It is very effective for These technologies use coefficient quantum
Coding, variable length coding, motion prediction and compensation, motion compensation and frame
Intra-frame coding integration and adaptive fields / frames
Including encoding. Motion prediction and compensation technology and motion compensation
1991, for the integrated technology of
U.S. Pat. No. 5,068,72 issued Nov. 26.
No.4 "Adaptive Motion Compensation for Digital Television"
Are described in greater detail and are hereby incorporated by reference.
See in. The circuit mechanism 150 that performs these functions is illustrated in FIG.
It is shown in FIG. The adaptive field / frame coding is 19
U.S. Pat. No. 5,091, U.S. Pat.
No. 782, Adaptive compression of continuous blocks of digital images
Apparatus and methods '', which is hereby incorporated by reference.
Incorporate here. US Special Issue issued March 3, 1992
No. 5,093,720 "Interlaced Digit
Motion Compensation of Digital Television Signals ", 1991/1
U.S. Pat. No. 5,057,910 issued on Jan. 15,
No. "Method of refreshing motion-compensated continuous video images
And devices ”are HDTV communications such as the system of the present invention.
Yet another motion compensation technique to help implement the communication system
Techniques, both of which are hereby incorporated by reference.
Incorporate in. Coefficient quantization is used to improve coding efficiency.
Second, the process of introducing small changes to the image. this
First weights each of the DCT coefficients, then decodes
Achieved by selecting 8 bits to transfer to the
You. The weight of each coefficient is fixed once assigned
Is not changed. Thus, for example, each coefficient is first set to 1
Expressed as a 2-bit number,
It can be divided by a weighting factor. However, the desired data
Another scaling is still required to achieve
It may be important. Therefore, the weighted coefficients are
Dividing by factor. Quantization factor is determined by quantization level
The quantization level is based on the scene complexity and perceptual characteristics.
And adjusted regularly. In a preferred embodiment of the invention, the quantity
Childing levels range from 0 to 31. Quantization level 0
When the quantization level is 30, the maximum precision is obtained.
It becomes small precision. Level 31 is secured and data is
Inform the decoder that it will not be sent. The 12-bit DCT coefficient is used as a weighting factor.
After scaling (normalizing) with both quantization factors,
Select the top 8 bits. In most cases, the upper 4 bits
Defaults to zero, so no information is lost. Only
When both the weighting factor and the quantization factor are small
Is used to prevent overflow and underflow
Need to clip the resulting coefficients
May be. The above-described quantization method uses the DC coefficient
Not applicable. Regardless of the quantization level, above the DC coefficient
The eight most significant bits are always selected. The quantization reduces the amplitude of the transform coefficient.
To improve compressibility
You. To take advantage of the result, a variable number of bits
An algorithm for assigning coefficients is needed. Variable length
The encoder uses a statistical coding technique, which is a quantization process.
Unlike information storage, the information is stored, so that the image does not deteriorate. In the preferred embodiment of the present invention, variable-length coding
Use Huffman coding. Huffman coding is theoretical
Approaching the entropy limit, beyond the probabilities of all possible events
Optimal Statistical Coding Method That Can Give Empirical Knowledge
well known. The encoder generates such a probability distribution
Before transferring any frames to the decoder.
Send. Next, the Huffman codeword is derived using this table.
You. The event with the highest probability of occurrence has a relatively short codeword
Assigned. The decoder also keeps the same codebook
It is possible to match each codeword to the actual event.
Wear. To keep the hardware simple,
Use fixed Huffman table generated based on range of materials
Then it is convenient. For Huffman coding, see above.
W. Pike's Digicipher-Fully Digital Communications
Road Compatibility HDTV Broadcasting System "
It is described in. The motion estimation and compensation subsystem of the present invention
150 first shows how the next frame will appear
To predict and then send the difference between the prediction and the actual image
Therefore, the image information is compressed. The correct prediction is just the previous
It is a frame. This type of time differential coding (DPCM)
Has little movement or spatial details
Works very well when there are few, but other
Is not very effective at times, and in some cases,
When the next frame is simply coded without prediction
(PCM) may be worse. Motion compensation is based on time pressure when motion occurs.
It is a means to improve the performance of the compacted structure. Apply motion compensation
To determine what has moved since the previous frame,
It is first necessary to determine where it has moved.
If this information is known to the encoder, the next frame
Move or displace the next frame that has not yet been transferred
System can be more accurately predicted. Encoder is a decoder
After playing back the same prediction, determine the difference between the prediction and the actual image.
Set. The motion matches the model used to predict the motion
And the motion prediction is accurate and the signal is noisy
If not, this error is actually zero. The displacement of the previous frame is determined by
It can be performed on a frame or pixel basis. Toes
Each frame, partial frame, or each pixel
A unique displacement (motion vector) can be generated
You. However, a single motion vector is generated per frame
To model a simple panning behavior of the whole image
Its usefulness is limited because
You. Ideally, generate a unique motion vector for each pixel.
You. However, motion estimation is a complex process,
Performs only at the encoder, as knowledge of the frame is required
Motion information per pixel that cannot be used by the decoder
The overhead involved in creating is excessive.
Therefore, the lateral dimensions and dimensions are equal to four DCT blocks.
"Super" with a vertical dimension equal to two DCT blocks
Select a partial area that is equal to
It is desirable to execute motion prediction in units. This size
Is four times the horizontal subsampling of the chrominance component.
Compatible with 2x vertical subsampling and same
Displace motion vector alone chrominance DCT block
It can be used to As shown in the encoder block diagram of FIG.
The motion compensation circuit 150 receives the output from the selector 137
With the configuration fed back to the input of DCT transform 134
Join. Similarly, the motion compensation of decoder 160 (FIG. 10)
162 is also installed at the output position of the inverse DCT transform. Image
Instead of tangent coding, motion compensation is used to predict the image.
Generate measurements first. Then, between this prediction and the actual image
Transform the difference, normalize the transform coefficients and statistically sign
Become The second of the two frames that guide motion estimation
Always before frames that appear after being reproduced by the decoder
It is a theme. Therefore, the encoder models the processing of the decoder.
I.e. a frame comparable to the decoder component 162
Includes delay and motion compensation circuits. As described above, the difference is calculated using motion compensation.
Instead of encoding the block directly into PCM
Bit rate may be lower due to
I'm right there. So get the smallest possible bit rate
The encoder needs the bits needed for each of the two methods
Number and then the minimum number of required bits per block
Choose a method. Required to inform the decoder of the choice.
It is important to note that the head is one bit per block. In an HDTV transmission system, a number of different
A television program is used as a common data stream for transmission.
Should be understood that they will be multiplexed together.
You. Each single channel image processor in the encoder is variable
Huffman coded data at the speed required for channel transmission
A rate buffer is installed to match the constant output rate.
Need to be prepared. This rate buffer is shown in FIG.
Can be implemented as a one-frame FIFO 140
it can. The total amount of storage in the FIFO is ± 1 image field.
It is large enough to handle fluctuations. Overflow of image output buffer FIFO
Or, to prevent underflow, use a FIFO input block.
It is necessary to constantly adjust the lock rate. this
Is achieved using a multiple quantization level coding structure
You. As the quantization level increases, the quantization becomes coarse and the
Lock is shorter, FIFO input block rate is higher
Become. Quantization level drops to minimum level of zero
And the quantization becomes denser, the blocks become longer and the FIFO
The input block rate will be lower. For this adjustment, the FIF
Essential to keep the bit rate to O relatively constant
effective. Buffer status is continuously monitored and stored
Blocks maintained within a given window
As long as the quantization level does not change. Low buffer level
Drop below threshold or above high threshold
Increase or decrease the quantization level, respectively.
I do. Underflow occurs during transmission of very simple images
Embedded bits into the channel to prevent
Can be The corresponding FIFO 164 is variable length
Equipped in the decoder (FIG. 10) before the decoder. Decoder
And a chrominance processor, generally designated 170.
Include and play the required RGB output. FIG. 11 is for transmitting compressed image data.
1 shows a basic communication system block used in the first embodiment. this
Are FEC encoding 180 and transmission filter 18 on the transmission side.
2, and QAM modulation. Dryness caused by transmission line
Interference and noise are indicated at 186 and 180, respectively. Receiving
On the side, demodulator 190, receive filter and adaptive quantization
A circuit 192, a tracking subsystem 194, and
An FEC decoder 196 is provided. Filter 182
And 192 are used for pulse shaping. Adaptive quantization
Is the reflection seen in typical VHF or UHF reception
Used to handle (multi-pass). The concatenated trellis coding and coding described in detail above.
And block coding structure to reduce the effects of transmission path errors.
And protect. Use two separate interleavers
A specific embodiment of the FEC encoder is shown in FIG. Also,
A corresponding FEC decoder embodiment is shown in FIG. In the figure
As shown, the FEC encoder is a Reed-Solomon encoder 20.
0 followed by a first interleaver 202,
Symbol generated by Reed-Solomon outer code
Are interleaved according to the first interleave format.
Behave. Interleaving is performed by trellis inner codes.
To distribute burst errors that may occur later.
Has the effect of Trellis encoder 204 provides I and Q signals.
The second component is output to the second interleaver 206, where
The encoded signal generated by the trellis inner code is
Interleave according to interleave format
You. This then occurs along the path of the encoded signal
This has the effect of dispersing burst errors that may occur. At the decoder, the coded signal is supplied to interleaver 2
10 which are processed by the
Recovered Reed-To Restore Solomon Symbol
To the trellis decoder 212. Trellis decoder
212 (for example, a coding rate of 3 /
In the case of 4, 32-QAM, the coding rate is 4/5).
Support the use of the inner code.
You. Reed-Solomo reconstructed by trellis decoder
The symbol is deinterleaved by another deinterleaver 214.
Input to the Reed-Solomon decoder 216
Is forced. The Reed-Solomon decoder 216 (for example,
For 16-QAM, coding rate 106/116, t =
5, 32-QAM, coding rate 145/155,
t = 5) is the burst generated by the trellis decoder
Since errors can be processed with the built-in burst error correction capability,
Used for side codes. For example, a least mean square (LMS) algorithm
The receiver can be equipped with an adaptive equalization circuit that uses
Wear. Such equalization circuits include noise, multipath, and
Signal constellation for soft decision in the presence of interference
The coefficient is constantly adjusted to optimize
Using a FIR (finite impulse response) filter
Can be configured. The adaptive equalization circuit uses NTSC interference
Image, color and audio carrier frequencies for improved rejection
Notch filtering of interfering NTSC signals automatically by number
Can be designed to do so. Thus, the present invention provides a compressed high-definition television
Digital signal with limited output and bandwidth, such as
Provide a practical system for transmission
Should be well understood. Based on the code of QPSK modulation
Modulation scheme based on QAM-based modulation system
Direct incorporation to form trellis coded QAM. this
Makes both bandwidth efficiency and data reliability both high and easy
A structure that can be realized is obtained. The present invention has been described with reference to specific embodiments.
However, those skilled in the art will appreciate that the invention
Many examples without departing from the spirit and scope of
It will be understood that adaptations and variations of are possible.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a QAM transmission system using concatenated coding. FIG. 2 is a block diagram of a trellis encoder according to the present invention. FIG. 3 is a block diagram of a trellis decoder according to the present invention. FIG. 4 shows a QAM constellation pattern divided into subsets according to the present invention. FIG. 5 is a chart defining the labeling of a subset of the constellation pattern of FIG. 4; 6 is a chart showing labeling of constellation points of the constellation pattern of FIG. 4; FIG. 7 is a graph comparing the performance of a concatenated coding scheme according to the present invention with that of a prior art coded QAM scheme. FIG. 8 is a block diagram of an HDTV communication system according to the present invention. FIG. 9 is a block diagram of a digital image encoder used in the system of FIG. FIG. 10 is a block diagram of a digital image decoder used in the system of FIG. 8; FIG. 11 shows forward error correction (FEC) according to the present invention.
FIG. 2 is a block diagram of a transmission system including encoding and QAM modulation / demodulation. FIG. 12 is a block diagram of an FEC encoder according to the present invention. FIG. 13 is a block diagram of an FEC decoder according to the present invention.

──────────────────────────────────────────────────の Continuing on the front page (51) Int.Cl. ⁷ Identification FI FI044 27/34 H04L 27/00 E (72) Inventor Scott A. Lely Hymetas Avenue, Lukadia, California, United States 1183 (72) Inventor Chris Hegard Ithaca, New York, USA, Earl Dee Number 2, Woodland Road 4 (72) Inventor Edward A. Klaus, Ariane Drive 2720-32, San Diego, California, USA (72) Inventor Gerold A. Heller Rancho Viejo Drive, Del Mar, California, United States 4932 (56) References JP-A-3-69295 (JP, A) JP-A-2-235491 ( JP, A) JP-A-2-195732 (JP, A) US Pat. No. 5,317,725 (US, A) European Patent 524625 (EP, B1) IEEE TRANSACTIONS ON BROADCASTING, VOL. 36, NO. 4, DECEMBER 1990 WOO PAIK "DIGIC IPHER (TM)-ALL DIGITAL, CHANNEL COMPAT IBLE, HDTV BROADCAS T SYSTEM" (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N ^7/ 13-7/24 H03M 7/30-13/00 H04L 27/34

Claims (1)

(57) Claims 1. A method for transmitting a digital television signal, the method comprising: dividing an image portion of the signal into blocks of PCM image data; Processing the blocks of PCM image data to generate DPCM data using the method; and selecting one of the PCM image data and DPCM data for transmission for each block based on predetermined criteria. Compressing the transformed data using a discrete cosine transform to generate a block of transform coefficients, quantizing the transform coefficients to improve their coding efficiency, and converting the transform coefficients after quantization into variable length codes. Encoding the variable length coded coefficients using a concatenated coding structure of a Reed-Solomon outer code and a trellis inner code, and transmitting Generating a coded signal, N point quadrature amplitude modulation having four subsets (QAM)
From the constellation pattern to quadrature amplitude modulation (QA
And transmitting the encoded signal as M) symbol consists city, each subset identified by a code word of different 2-bit
Separated and said N-point QAM constellation pattern
, And the trellis inner code is the Reed-Solomon outer code.
Code the symbol from the symbol, the first bit of the symbol
Process with a binary convolutional coding algorithm with a half rate
Encoded by the constellation pattern
Assigned to the subset in which the symbol exists.
Provide a two-bit codeword , said two-bit codeword being used to generate a modulation function.
The remaining bits of the symbol are mapped with the remaining bits of the symbol, the remaining bits of the symbol being defined by the codeword.
Of N / 4 symbol points included in the defined subset
In one aspect, the method comprises associating the symbols . 2. The concatenated coding structure interleaves the symbols generated by the Reed-Solomon outer code according to a first interleave format, and then distributes burst errors that may be generated by a trellis inner code. Interleaving the coded signal generated by the trellis inner code according to a second interleaving format, and then dispersing burst errors that may be generated along the transmission path of the coded signal. The method of claim 1, comprising: 3. The method of claim 2, wherein the two-bit codeword comprises the least significant bit of the modulation function, and wherein the remaining bits comprise the most significant bit of the modulation function. The method of claim 1 . Wherein said method further includes the steps of receiving at the receiver the carrier, by demodulating the received carrier at said receiver, comprising the steps of restoring the modulation function, the recovered modulation function, Deinterleaving using the inverse of the second interleaving format; pruning the deinterleaved modulation function to generate a set of metrics corresponding to the subset and identified by the remaining bits. Generating a number of bytes representing various conditional decisions of the signal points; and using the metric in a trellis code algorithm to decode a rate 1/2 binary convolutional code,
Restoring the bits; encoding the restored first bits using a binary convolutional coding algorithm with a coding rate of ， to reproduce the codeword; Responsively, selecting one of the conditional decision bytes; combining the selected byte with a reconstructed first bit to generate a decoded output; a step of deinterleaving using reverse of the first interleave format, characterized in that it consists method of claim 1, wherein. 5. The method of claim 4 , further comprising the step of: decoding the deinterleaved decoded output using a Reed-Solomon symbol error correction decoding algorithm. 6. The method of claim 1, wherein the Reed-Solomon decoding algorithm recovers variable length coding coefficients from the deinterleaved decoded output , and the method further comprises decoding the recovered variable length coding coefficients. Restoring the transform coefficients representing the image portion, and inversely transforming the restored transform coefficients to obtain PCM and DPC
Restoring image data according to at least one of the M formats; processing the restored DPCM data by motion compensation;
6. The method of claim 5 , further comprising the steps of: restoring PCM image data represented by the restored DPCM data; and formatting the restored PCM image data for output to a digital television receiver. the method of. 7. Quadrature amplitude modulation ( QA) including compressed image information.
A method of decoding a digital television signal transmitted using the M), said method comprising the steps of: receiving a carrier containing said signals, demodulates the received carrier wave, interleaved containing the compressed image information Reconstructing the modulation function, deinterleaving the reconstructed modulation function, decoding the deinterleaved modulation function with an inner trellis decoding algorithm of a concatenated decoder, and displaying an interleaved read representing compressed image information. Reconstructing the Solomon symbol, deinterleaving the reconstructed Reed-Solomon symbol, and inputting the reed to the outer Reed-Solomon decoding algorithm of the concatenated decoder to provide
Reconstructing a variable-length coding coefficient from a Solomon symbol; decoding the reconstructed variable-length coding coefficient to restore a transform coefficient representing the compressed image information; Converting and restoring image data according to at least one of the PCM and DPCM formats; processing the restored DPCM data by motion compensation;
Reconstructing PCM image data represented by reconstructed DPCM data; and formatting the reconstructed PCM image data for output to a digital television receiver , wherein the modulation function comprises two bits. Multiple QAM controllers with codewords
Identify one of the stellation subsets and
A signal point in the one subset is represented by an n-2 bit part.
Comprising an N-point QAM modulation function, wherein the method further comprises pruning the reconstructed modulation function to form a pair with the subset.
Generate a corresponding set of metrics and n-2 bits
Multiple conditional decisions of signal points identified by
Stage for generating a large number of n-2 bit subgroups to represent
Floor and using the metric in the trellis decoding algorithm
To decode a binary convolutional code having a coding rate of 、,
Restoring the bits, and converting the restored first bits to a binary convolutional code having a code rate of ２.
Decoding with an encoding algorithm and reproducing the codeword
And in response to the reproduced codeword, the plurality of n-2
Selecting one of the sub-groups of the bits and combining the selected sub-group with the recovered first bit
And generating a Reed-Solomon symbol . 8. The modulation function according to claim 2, wherein said two-bit codeword comprises the least significant bit of said modulation function, and said remaining bits comprise the most significant bit of said modulation function.
The method of claim 7 . 9. An apparatus for transmitting digital television signals, the apparatus comprising: means for processing blocks of PCM image data using motion estimation and compensation to generate corresponding DPCM data; Is compared with the corresponding DPCM data generated by the processing means, and the PCM image data and the DPC
Means for selecting one of the M data for transmission, means for compressing the selected data by a discrete cosine transform to provide a block of transform coefficients, and quantizing the block of transform coefficients to reduce their coding efficiency. means for improved, coupled to an output of said quantizing means, and means for variable length encoding the transform coefficients, lead - include Solomon outer coder and a trellis inner coder, which is the variable length coding transform Means for encoding the coefficients to provide an encoded signal for transmission; N-point quadrature amplitude modulation (QAM) having four subsets
Quadrature amplitude modulation (QA) from constellation pattern
Made, and transmitting the encoded signal as M) symbols, each subset identified by a code word of different 2-bit
Separated and said N-point QAM constellation pattern
, And the trellis inner encoder includes the Reed-Solomon outer
Replace the symbol from the encoder with the first bit of the symbol
Processed by a binary convolutional coding algorithm with a coding rate of 1/2
The constellation path
Assign to the subset where the symbol exists in the turn
To provide a two-bit codeword , said two-bit codeword being used to generate a modulation function.
Mapped with the remaining bits of the symbol
The bits are sub-bits defined by the 2-bit codeword.
One of the N / 4 symbol points included in the set
Associated with the symbol, the transmitting means modulates a carrier using the modulation function.
Rukoto, apparatus according to claim. 10. The apparatus, said coupling encoder, the lead - the symbols generated by Solomon outer coder and interleaved with the first interleave format, the possibility of a subsequently generated by the trellis inner coder A first interleaver for dispersing a certain burst error, and an encoded signal generated by the trellis inner encoder may be interleaved in a second interleave format, and then generated along a transmission path of the encoded signal. 10. The apparatus of claim 9 , further comprising: a second interleaver for dispersing burst errors of interest. 11. The modulation function of claim 2, wherein said two-bit codeword comprises a least significant bit of said modulation function, and said remaining bits comprise a most significant bit of said modulation function. An apparatus according to claim 9 . 12. The apparatus further comprising: means for modulating the carrier at a receiver to recover the modulation function; means for deinterleaving the recovered modulation function using an inverse of the second interleaving format. Means for pruning the deinterleaved modulation function to generate a set of metrics corresponding to the subset and generating various conditional decisions of signal points identified by the remaining bits; Concatenated to receive the metric, code rate 1
Trellis decoding means for decoding the binary convolutional code of / 2 and restoring the first bit, encoding the restored first bit by a binary convolutional encoding algorithm having a coding rate of 1/2, Means for reproducing the codeword; means for selecting one of the conditional decision bytes in response to the reproduced codeword; combining the selected byte with the recovered first bit. The apparatus of claim 9 , further comprising: means for generating a decoded output; and means for deinterleaving the decoded output using an inverse of the first interleaving format. Wherein said apparatus further decoded output deinterleaved, Reed - using Solomon symbol error correcting decoding algorithm, characterized in that it comprises means for decoding apparatus of claim 12, wherein. 14. The apparatus, said lead - Solomon decoding algorithm recovers variable length coded coefficients from the decoded output which is de-interleaved, the apparatus further decoding the reconstructed variable length coded coefficient Means for restoring a transform coefficient, and inversely transforming the restored transform coefficient to obtain a PCM and a DPC
Means for restoring image data in at least one of the M formats; processing the restored DPCM data by motion compensation;
Means for restoring PCM image data represented by restored DPCM data; means for formatting the restored PCM image data from the restored DPCM data processing means and the inverse transform means for output to a digital television receiver. characterized in that it consists of, the method of claim 13. 15. The method according to claim 14, further comprising a quadrature amplitude modulation (Q)
AM) for decoding a digital television signal transmitted using AM , the apparatus receiving a carrier wave containing the digital television signal.
Means that demodulates the carrier including the signal, QAM recovery to restore the interleaved modulation function containing the compressed image information
A regulator unit, a first deinterleaver for deinterleaving the recovered modulation function, a modulation function which is de-interleaved, inner trellis decoding
And decoded using vessels, compressed image information interleaved Reed represent - a connecting decoder to recover the Solomon symbols, de-interleaved Reed - restoring the variable length coding coefficients from Solomon symbol, the connecting decoder A second deinterleaver for deinterleaving the reconstructed Reed-Solomon symbols for input to the outer Reed-Solomon decoder, and decoding the reconstructed variable length coded coefficients to produce the compressed image. Means for restoring transform coefficients representing image information; means for inversely transforming transform coefficients to restore image data in at least one of the PCM and DPCM formats; and processing the restored DPCM data by motion compensation. Means for restoring PCM image data represented by restored DPCM data; It restored PC from the means
Means for formatting the M image data for output to a digital television receiver , said modulation function comprising:
Identify one of the stellation subsets and
A signal point in the one subset is represented by an n-2 bit part.
Comprising an N-point QAM modulation function, said apparatus further pruning the reconstructed modulation function and pairing with said subset.
Generate a corresponding set of metrics and n-2 bits
Multiple conditional decisions of signal points identified by
A method for generating a large number of n-2 bit subgroups to represent
Stage and the reconstructed first bit are binary convolved at a code rate of １ ／
Encode with an encoding algorithm and reproduce the codeword
Means, and in response to the reconstructed encoder, the plurality of n-2 decoders.
Means for selecting one of the sub-groups of bits, combining the selected sub-group with the recovered first bit
Means for generating Reed-Solomon symbols , wherein said concatenated decoder is adapted for use in said inner trellis decoder.
In order to receive the metric concatenated and code rate
The first bit is decoded by decoding a half binary convolutional code.
To source, and wherein the. 17. The device further characterized in that it consists of the adaptive equalization circuit coupled between the coupling decoder to the QAM demodulator apparatus of claim 16, wherein.